France summons Elon Musk in X probe over deepfakes and child abuse material

French prosecutors have summoned Elon Musk to Paris for a voluntary interview in the case against X. Former platform chief Linda Yaccarino has also been invited to be questioned, while company employees are being interviewed this week as witnesses.

Why leadership was summoned

The investigation is led by the cybercrime unit of the Paris prosecutor's office. Formally, this is not yet about filing charges but about voluntary interviews: investigators want the position of X's leadership as the people who ran the platform when, according to the prosecution, the disputed events took place. French authorities also stress that even a possible no-show will not stop the case.

Musk is summoned not as an outside observer but as the owner and de facto manager of a service that operates in France and is obliged to comply with local law. At the same time, authorities are collecting testimony from X employees in France.

This format shows that prosecutors are looking not at a single post or one specific account, but at the platform's internal processes: how moderation is structured, which decisions were made about algorithms, how complaints were handled and what measures the company took after the scandals around Grok and illegal content.

What investigators are examining

Prosecutors are pursuing several lines of inquiry. According to case materials and French and international media reports, the investigation covers not only sexualized deepfakes but a broader set of complaints against X and its associated chatbot, Grok.

• alleged complicity in storing and distributing images of sexualized child abuse
• generating and distributing sexualized deepfakes without the consent of the people depicted
• possible manipulation of an automated system through changes to recommendation algorithms
• Grok posts that French authorities link to the denial of crimes against humanity
• compliance measures: how X detects violations and what it does after complaints

The French prosecutor's office stresses that the goal of the current stage is to hear management's explanations and understand which measures X is prepared to implement.

"These interviews should give leadership the opportunity to set out

its position."

The company itself has previously denied wrongdoing and called the investigation politicized.

How the case grew

The story did not begin with deepfakes. The first reports reached prosecutors on 12 January 2025: French lawmaker Eric Bothorel and an official working on cybercrime said that changes to X's recommendation algorithms could have distorted the platform's operation. In summer 2025, prosecutors opened a preliminary investigation, and in February 2026 French law enforcement searched X's Paris office with Europol's support.

The focus then widened. According to Le Monde, investigators are separately examining how X changed its tools for detecting child sexual abuse material: the platform dropped Thorn's Safer system in favor of its own solution, after which, according to prosecutors, the number of reports from France to the US center NCMEC fell by 81.4% between June and October 2025. Against this backdrop, the Grok scandal erupted: regulators and researchers say the chatbot generated sexualized deepfakes of women and minors from simple text prompts.

The international context adds pressure. The EU and the UK have already run separate checks on X and xAI over deepfakes, personal data processing and moderation mechanisms. For France, the case has become a test of how firmly the state is prepared to apply criminal law to global digital platforms when, in the view of investigators, their products, algorithms or safeguards systematically facilitate violations.

What this means

The X case shows that complaints about AI services quickly move from content moderation into criminal law. If the French investigation goes further, it will signal to X, xAI and other platforms that liability is now assessed not only by users' posts but also by product architecture, algorithm configuration and the quality of safeguards.